CCRL Discussion Board

When you are not in check, illegal moves come in two kinds: the King stumbling into an attacked square, and moving a pinned piece. So when all btm positions contain the pseudo-legal move count, woud scan the squares coverd by the white pieces (by generating white moves). In a first iteration, I woul...

Well, nothing of the sort has been agreed by me.. Building WDL would take me exactly as long as building DTZ, so building them first would be a total waste of time. The reason this did not work for Thompson was that he was building EGTs on disk, and that the process was totally dominated by disk I/O...

This was long ago, as in the mean time I have been distracted by many other Chess-programming projects. (Butthis one is still on the to-do list!). So let me see if I can still answer that: The described process indeed counts pseudo-legal moves, and then subtract the illegal moves later, for all posi...

Well, my engines have no opening book. It usually depends on the engine, each engine has its own format. But I recently added probing code for a GUI book to WinBoard, using the Polyglot book format, so I know now how that works. It seems to be quite a sensible format: it stores (position, move, weig...

Indeed from an implementation point of view a GUI opening book is exactly the same as a GUI tablebase. But due to the nature of the information they contain, the demands are different. With openings, you start in the book, and when you are out, you are out of it. There is no big advantage for engine...

If GUIs support tablebase probing (I do not know anything specifically about Arena, so I am talking hypothetically) they can do it only to bypass the engine in cases where the current board position is actually inside the tablebase. This is different from an engine using tablebases, as in the latter...

Well, in general, it is. For end-games where this strategy could require over 50 moves, you can always use DTZ. I know little about KNNKP, but it hardly seems it could be a problem there: if you sac a knight, the only way I can imagine to win the remaining KNKP is if the black King is trapped in a c...

In general I would want the quickest conversion. But I am starting to realize now that this only applies to conversions that capture an opponent piece or move my own Pawns towards promotion. Not to forcing a conversion by senseless sacrifice of my own material, just for the sake of having a conversi...

Let me think aloud about this: When all won(<N ply) positions have been identified, and used to verify all lost(N+1 ply) positions (i.e. this cycle is complete), leap-frog can immediately start working on the currently cached slice for identifying a partial set of won(N+1 ply) positions (accessible ...

I was busy making a new XBoard release. But that is out now.

I succeeded in installing an Ubuntu on the so far unused disk of my C2D, after some initial problems with the display. I tested the tranfer rate as you described, and I get a speed or 58MB/sec on /dev/sda2, and 62MB/sec on /dev/sdb3 (the partitions I reserved for EGTB generation), reading 8GB at the...

OK, thanks. I am getting more and more convinced that it is essential to do this under Linux in order to get acceptable performance. Until recently, I had no experience with Linux whatsoever, though. So I am learning now how to do elementary things, like editing and compiling. This is why I had a sl...

OK, I tried this on my laptop, (currently the only computer where I have Linux installed), for 1K blocks of 8M, and it was reaching a spead of 31MB/s. This should be somewhere halfway the disk. (I have a 120GB disk, and the irst 40GB consists of Windows partitions. I tried this on /dev/sda7, which i...

How exactly do you measure the speed of your hard disk? Create a file far larger than your memory, and then read it sequentially? And especially for writing, do you write in an existing file? I did a benchmark under Windows XP on my laptop, alternating seeks in an existing file on an NTFS partition ...

On my 2.4GHz Core 2 Duo I can do 900MB/sec in dirty cache replacements (i.e. one read + one write). Read only would give 3.5GB/sec. On a hard disk 50MB/sec read or write would already be a good speed. So there is a significant factor between the two, and what you suggest is certainly worthwile. In t...

Hmm... So I have to store this bitbase too (at least in endgames where black does have more than just a king). I am hesitant to store this in the dtm-byte, because in 6 men endgames the remaining 7 bits are not enough to store dtm's that are >127. Maybe I have to make the datastructure for 3,4,5 me...

Getting confused... How can I generate DTM=0 entries in a wtm-file?? they do not exist! So how can you store any DTM then? I am not sure what you mean here. In principle, the side to move is one of the chararacteristics identifying the position. If pieces are on the same squares, but the other side...

For complicated positions, wouldn't it be more efficient to generate both at once? One idea I had was using negative numbers for losses, 0 fr a draw, and positive numbers for wins You can do that if you store the actual DTM, and have enough bits available to split the coding space in sets for white...

It seems you are generating wins for both sides at the same time, and do not distinguish wins for black and wins for white. White should go for the shortest mate-in-n that is a win for white, (making the predecessor a mate-in-n+1), and if thee is none, to an undecided position (in which case the pos...

I posted the more advanced version of my 5-men EGTB generator now on my website as well: http://home.hccnet.nl/h.g.muller/5men13.html On my 2.4 GHz Core 2 Duo (E6600) it takes 86.5 sec to generate KBBKN, and 141.5 sec for KQQKQ with this program. The latter end-game is the one that takes the longest...

It is complicated, I almost get it. I will try to compile your C-program from your website to check the results. It's funny to see that it's possible in about 500 lines code. I think I have to forget about my standard way of generating moves etc. which I use for my engine. It seems handy but genera...

Ok about this symmetry. I want to ask a questions to see if I understand it: Let call a position this is in the egtb "RP" (Real Position). And a position that is not (so a position that has to be rotated and mirrored to do a probe) "VP" (Virtual Position) 1) I'm testing a RP whe...

Pehaps you should try first without diagonal symmetry (or perhaps without any symmetry at all), to see if you can reproduce the numbers of KRK for the various DTM (e.g. the numbers you get from the generator on my website). If these match, you know that the basic logic is OK, and you can switch on t...

If you generate the DTx in increasing order, you can always label the predecessors of a DTx=2 btm position as DTx=3 (wtm). Because you know that you have not assigned DTx > 3 to any position yet. Otherwise, you would have to check in order to avoid make an already assigned DTx worse. For wtm positio...

You touch on an interesting subject, which I often wondered about. The DTM or DTC are an absolute measure for the winning distance, but very erratically related to the actual position. This makes it impossible for a Human to remember them, and makes it very difficult for a computer program to develo...